Call detail record (CDR) data from mobile communication carriers offer an emerging and
promising source of information for analysis of traffic problems. To date, research on
insights and information to be gleaned from CDR data for transportation analysis has been
slow, and there has been little progress on development of specific applications. This paper
proposes the traffic semantic concept to extract traffic commuters’ origins and destinations
information from the mobile phone CDR data and then use the extracted data for traffic
zone division. A K-means clustering method was used to classify a cell-area (the area covered
by a base stations) and tag a certain land use category or traffic semantic attribute
(such as working, residential, or urban road) based on four feature data (including
real-time user volume, inflow, outflow, and incremental flow) extracted from the CDR data.
By combining the geographic information of mobile phone base stations, the roadway network
within Beijing’s Sixth Ring Road was divided into a total of 73 traffic zones using
another K-means clustering algorithm. Additionally, we proposed a traffic zone
attribute-index to measure tendency of traffic zones to be residential or working. The calculated
attribute-index values of 73 traffic zones in Beijing were consistent with the actual
traffic and land-use data. The case study demonstrates that effective traffic and travel data
can be obtained from mobile phones as portable sensors and base stations as fixed sensors,
providing an opportunity to improve the analysis of complex travel patterns and behaviors
for travel demand modeling and transportation planning.